In many application fields such as disaster management or personal planning, efficient mobile services should capture users' context in order to provide them with the most appropriate information at a given time. Situation-based services aim towards this direction. They go beyond location-based services as they take a user's current situation into account. Situations include user profiles together with dimensions such as their location or current activity. Depending on the application but also within the same application the time interval of interest can be of various granularities. It is crucial to handle time granularity efficiently and in a well-defined manner. In this paper, we investigate the handling of various granularities for situations and situation sequences through coarsening methods and we propose an algorithm that we used in concrete projects.